Nicotine & Tobacco Research, Volume 16, Supplement 1 (January 2014) S44–S55

Original Investigation

Who Quits? An Overview of Quitters in Low- and Middle-Income Countries Ce Shang PhD1, Frank J. Chaloupka PhD1,2, Deliana Kostova PhD3 1Health Policy Center, Institute for Health Research and Policy, University of Illinois at Chicago, Chicago, IL; 2Department of Economics, University of Illinois at Chicago, Chicago, IL; 3Centers for Disease Control and Prevention, Atlanta, GA

Corresponding Author: Ce Shang, PhD, Health Policy Center, Institute for Health Research and Policy, University of Illinois at Chicago, Room 422, 1747 W. Roosevelt Road, Chicago, IL 60608, USA. Telephone: 312-996-0774; Fax: 312-996-2703; E-mail: [email protected] Received April 29, 2013; accepted October 8, 2013

Introduction: Using the Global Adult Tobacco Surveys from 14 primarily low- and middle-income countries, we describe the association between the probability of being a recent quitter and a number of demographic and policy-relevant factors such as exposure to warning labels, work-site smoking bans, antismoking media messaging, tobacco marketing, and current cigarette and bidi prices. Methods: Logistic regressions were used to examine the potential correlates of recent quitting and recent quit attempts. Results: After accounting for country-specific attributes in pooled analyses, we found that higher rates of exposure to worksite smoking bans are associated with higher odds of being a quitter (odds ratio [OR] with 95% confidence interval [CI] = 1.13 [1.04, 1.22]). Exposure to antismoking media messaging (OR with 95% CI = 1.08 [1.00, 1.17]), work-site smoking bans (OR with 95% CI = 1.11 [0.99, 1.26]), and warning labels (OR with 95% CI = 1.03 [1.01, 1.05]); cigarette prices (OR with 95% CI = 1.01 [1.00, 1.02]); and bidi prices (OR with 95% CI =1.17 [1.11, 1.22]) are factors associated with higher odds of recent quit attempts in the pooled analysis. These effects vary by country. Exposure to warning labels is found to be associated with greater likelihood of recent quitting in Egypt (OR with 95% CI = 3.20 [1.53, 6.68]), and the positive association between exposure to work-site smoking bans and quitting is particularly strong for Southeast Asia (OR with 95% CI = 1.20 [1.06, 1.35]) and Asia Pacific countries (OR with 95% CI = 1.85 [0.93, 3.68]). Additionally, exposure to tobacco industry marketing is significantly associated with smaller odds of quitting in Asia Pacific (OR with 95% CI = 0.83 [0.79, 0.87]) and Latin American countries (OR with 95% CI = 0.78 [0.74, 0.82]). Conclusions: Although our results vary by country, they generally suggest that greater exposure to tobacco control polices is significantly associated with quitting.

Introduction Tobacco use is one of the leading causes of preventable death worldwide. Although the majority of the world’s smokers reside in low- and middle-income countries (LMICs), the quit rate among smokers in LMICs is relatively low (Jha et  al., 2008; Rani, Bonu, Jha, Nguyen, & Jamjoum, 2003). To address this, the World Health Organization (WHO) has identified tobacco cessation as a major goal in its published guidelines of the Framework Convention on Tobacco Control 2006 (WHO FCTC) (WHO, 2010b). Although smoking cessation is recognized as an important aspect of tobacco control in LMICs, less is known about the factors linked to cessation in these countries. Using individuallevel data on smokers from 13 LMICs and Poland obtained from the Global Adult Tobacco Survey (GATS) 2008–2010, we examine potential correlates of quitting within and across

this group of countries. (Countries were classified into income groups according to their 2011 per capita gross national income following the World Bank Atlas method. Countries are classified as high income if they have a gross national income per capita of $12,476 or more. Poland has a gross national income per capita of $12,480 which is slightly above the cutoff. Therefore, we included Poland in our analyses alongside 13 LMICs.) We describe the association between recent quitting and a comprehensive set of policy-relevant and individual-specific factors. Using individual GATS responses, we construct a number of location-specific index variables, which reflect the local prevalence of work-site smoking bans, cigarette warning labels, tobacco advertising, tobacco promotion, antismoking information, and prices paid for cigarettes (and bidis for India and Bangladesh). By exploring these factors as correlates of quitting, we evaluate their potential as cessation-promoting mechanisms among smokers in LMICs.

doi:10.1093/ntr/ntt179 Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco 2013. This work is written by (a) US Government employee(s) and is in the public domain in the US.

S44

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

Abstract

Nicotine & Tobacco Research, Volume 16, Supplement 1 (January 2014)

Data and Methods The GATS is an ongoing nationally representative household survey of adults aged 15 years or older, which has been conducted in 14 countries between 2008 and 2010. It collects information on respondents’ demographic characteristics, tobacco use, exposure to tobacco control policies, and tobacco marketing. In GATS, respondents who identify themselves as past smokers are asked to report how long it had been since they quit smoking. This allows us to describe measures of quitting within the past 12  months, which can be evaluated in the context of recent exposure to tobacco control policies and tobacco marketing, as well as to demographic characteristics. These contemporaneous quitting measures include an indicator of respondents who quit in the past 12  months (the ratio of the number of

respondents who quit in the past 12 months to the number of smokers 12 months ago) and an indicator of smokers who made at least one unsuccessful quit attempt (the ratio of the number of current smokers who attempted to quit in the past 12 months to the number of current smokers). The sample of smokers 12  months ago consists of both current smokers and those who quit in the past 12 months, and that the combined quitting and quit attempt rates can be derived as a weighted average of the two indicators. We present these quitting measures along with smoking prevalence in the studied countries in Figure 1. Countries with relatively high smoking prevalence rates such as China and Russia tend to have relatively low recent quit rates, whereas Latin American countries have the highest quit rates. The GATS asks smokers to report expenditures on their last purchase of cigarettes (and bidis for respondents from India and Bangladesh), as well as the number of sticks purchased. Using this information, the price paid per cigarette (bidi) can be derived for each smoker. Since individual-level prices and individual smoking intensity are likely to be simultaneously determined (heavier smokers are more likely to seek out lower prices while lower prices encourage heavier smoking), individual-level prices would be endogenous in models of smoking cessation. To address this simultaneity bias, our analyses use market-level prices, derived as the primary sampling unit (PSU)-specific consumption weighted average cigarette (bidi) price paid per 20 sticks. This approach has been detailed in the International Agency for Research on Cancer (IARC) Handbook (IARC Handbooks of Cancer Prevention, Tobacco Control, 2008) and Economics of Tobacco Toolkit (WHO, 2010a). In order to make prices comparable across countries, we convert them into a common international dollar currency using purchasing power parity adjustment factors, and then into constant 2010 international dollars using the index of average consumer prices published by the International Monetary Fund (Table 1). Individual-level demographic controls include age, gender, education, rural residence, wealth, household size and occupation type (Tables 1 and 2). Age is defined by binary indicator variables for four age categories (15–24, 25–39, 40–64, 65 and older). Education level is described through five categories: no education/less than primary, primary, secondary, high school, and college or higher. Occupation type is described through three categories: indoor, outdoor, and unemployed/unspecified. Wealth is measured from survey questions that inquire about the possession of certain personal and household items (electricity, flush toilet, fixed telephone, cellular phone, television, radio, refrigerator, car/bike/boat, moped/scooter/motorcycle, washing machine, and any other surveyed assets) and is defined as the fraction of surveyed items, which the respondent has in their possession, weighted by the per capita gross domestic product of the respondent’s country. Besides individual demographic controls, our study employs a number of indices constructed from the individual responses of GATS participants. These include exposure to tobacco control policies, exposure to antismoking media messaging, and exposure to tobacco marketing (Tables 1 and 2). These indices are constructed as PSU-level aggregates, which has a number of advantages over using the underlying individual-level exposure status. First, individual exposure may have a reverse causality link to quitting behavior—for instance, antismoking messaging may target and be observed disproportionately more by the type of person who is more prone to quitting in the first place or tobacco marketing may be disproportionately targeting

S45

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

Over the past decade, a growing body of research has focused on examining the impacts of price and tobacco control policies on tobacco use in LMICs, building upon the evidence demonstrating the effectiveness of these interventions in high-income countries (HICs) (Chaloupka et al., 2011; Chaloupka & Warner, 2000; Guindon, Perucic, & Boisclair, 2003; Kostova, Ross, Blecher, & Markowitz, 2011; Ranson et  al., 2002). However, there are few studies that examine the factors in promoting smoking cessation in LMICs. Kostova (2012) examines the impact of price on smoking transitions before the age of 15 in a set of 40 LMICs and finds that in early adolescence, prices are more effective in driving initiation rather than cessation. Ross et al. (in press) find that higher taxes have effectively increased cessation among adults in three Eastern European countries during the transitional period of the 1990s and 2000s. Existing studies of U.S. data have estimated that the short-run price elasticity of smoking cessation, which reflects smokers’ initial efforts to quit in response to price, increases ranges from 0.3 to 0.9 and falls in the longer run as some quitters relapse (DeCicca, Kenkel, & Mathios, 2008; Tauras, 2004). Graphic warning labels and mass-media antismoking campaigns that encourage quitting have been shown to play a role in increasing interest in cessation and quit attempts in HICs (Centers for Disease Control and Prevention, 2012; Davis, Nonnemaker, Farrelly, & Niederdeppe, 2011; Hammond, Fong, McNeill, Borland, & Cummings, 2006). However, the effect of these nonfiscal approaches on cessation in LMICs has not been extensively studied. The relationship between cessation and tobacco control measures such as taxes in LMICs is likely to be different from that in HICs. On the one hand, complicated/tiered tax structures in many LMICs can widen the range of cigarette prices within countries (Chaloupka, Kostova, & Shang, 2013; Shang, Chaloupka, Fong, & Zahra, 2013). This provides an incentive for smokers to switch between cigarette brands or tobacco products in response to higher taxes, potentially reducing the full impact of tax increases on lowering prevalence and consumption. Similarly, economic growth in some LMICs that results in significant income increases can make tobacco products more affordable, encouraging further shifts in consumption (Blecher & van Walbeek, 2004, 2009; Kostova et al., 2012). On the other hand, given the relatively low awareness of tobacco health risks in some LMICs (King et al., 2010), informational policy tools such as graphic warning labels and mass media antismoking campaigns may have a relatively larger impact on cessation in LMICs.

Quitters in low- and middle-income countries the tobacco promotion index. The tobacco advertising index was constructed using a similar formula: first, we estimated the fraction of advertising outlets (stores, television, newspapers or magazines, and any other outlets) that have exposed each respondent to tobacco advertising in the past 30 days, then averaged these fractions at the PSU level and scaled them from 1 to 10 to produce the tobacco advertising index. Logistic regressions were used to examine the potential correlates of recent quitting (the probability of quitting in the past 12  months) and recent quit attempts (unsuccessful quit attempts in the past 12  months). For individual country estimates, the standard errors were clustered at the PSU level; in the pooled analysis, they were clustered at the country level. All models include individual demographic controls for age, gender, education, wealth, household size, rural residence, occupation type, and PSU-level indices for antismoking information, tobacco promotion, tobacco advertising, and warning labels, and prices paid for cigarettes. Bidi prices were included in models for India and Bangladesh. To examine how workplace smoking bans may impact quitting depending on the type of employment, we included interaction terms between the workplace smoking ban index and the indicators for indoor and “other” occupation (“other” refers to outdoor and unemployed/ unspecified occupation). Pooled models include country fixed effects to control for unmeasured country-specific factors that may impact cessation behavior. The analysis sample for the models of recent quitting consists of current smokers and those who quit in the past year and the sample for models of recent quit attempts consists of current smokers only.

Figure 1.  Quitting and quit attempts in the past 12 months and smoking prevalence by country

S46

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

individuals who are committed smokers and less likely to quit in the first place. The reverse causality bias is reduced by aggregating individual exposure responses at the PSU level. Second, aggregating individual exposure responses at the PSU level reduces misclassification bias occurring when some respondents underreport and some overreport personal exposure. Third, the PSU-level indices may also capture subnational differences in policy implementation and enforcement. They are defined as follows. The work-site smoking ban index is constructed as the PSU-level average of individual responses that designate the presence of smoking restrictions at the respondent’s work-site (0, no restriction; 1, some restriction; 2, full restriction). The warning label index is constructed at the PSU level as the fraction of respondents who report noticing warning labels among those who have seen cigarette packs in the past 30 days, scaled from 1 to 10. For developing the antismoking information index, we first estimated, for each respondent, the fraction of media outlets (newspapers or magazines, television, radio, billboard, and any other outlets) that have exposed the respondent to antismoking information in the past 30 days. These individual-specific fractions were then averaged across all respondents in a PSU and scaled from 1 to 10 to produce the index. For developing the tobacco promotion index, we first estimated, for each respondent, the fraction of promotion approaches (free samples, clothing with brand names, and any other approaches) that have been observed by the respondent in the past 30 days. Similarly to the antismoking information index, these individual-specific fractions were then averaged across respondents in a PSU and scaled from 1 to 10 to produce

Nicotine & Tobacco Research, Volume 16, Supplement 1 (January 2014) Table 1.  Variable Descriptions and Definitions Individual-level variables   Quit in past 12 months   Quit attempt in past 12 months  Age  Education  Wealth

       

Household size Rural residence Indoor occupation Outdoor occupation

Indicator equals 1 if the respondent has quit smoking in the past 12 months, 0 if smokes at the time of survey Indicator equals 1 if the smoker at the time of survey attempted and failed to quit in the past 12 months, 0 otherwise Binary indicators for four age categories: 15–24, 25–39, 40–64, 65+ Binary indicators for five education categories: no education/less than primary, primary, secondary, high school, college or higher The fraction of GATS-surveyed household items (electricity, flush toilet, and any other surveyed assets) that the respondents has in their possession, weighted by the per capita gross domestic product of the respondent’s country Number of household members Indicator equals 1 if the respondent lives in rural area, 0 otherwise Indicator equals 1 if the respondent works indoors, 0 otherwise Indicator equals 1 if the respondent works outdoors, 0 otherwise

  Work-site smoking ban index

  Warning label index   Antismoking information indexa

  Tobacco promotion indexa

  Tobacco advertising indexa

  Cigarette price   Bidi price

The average of individual responses that designate the presence of smoking restrictions at the respondent’s work-site (0 = no restriction, 1 = some restriction, 2 = full restriction) The fraction of respondents who report noticing warning labels among those who have seen cigarette packs in the past 30 days, scaled from 1 to 10 Out of a number of possible antismoking media outlets (newspapers or magazines, television, radio, billboard, and any other outlets), the fraction that each respondent was exposed to in the past 30 days was calculated. These individual-specific fractions were then averaged and scaled from 1 to 10. Out of a number of possible promotion approaches (free samples, clothing with brand names, and any other approaches), the fraction that each respondent observed in the past 30 days was calculated. These individual-specific fractions were then averaged and scaled from 1 to 10. Out of a number of possible advertising outlets (stores, television, newspapers or magazines, and any other outlets), the fraction that each respondent was exposed to tobacco advertising in the past 30 days was calculated. These individual-specific fractions were then averaged and scaled from 1 to 10. Average price paid per 20 cigarettes in constant 2010 international dollars Average price paid per 20 bidis in constant 2010 international dollars

Note. GATS = Global Adult Tobacco Survey; PSU = primary sampling unit. aThe indices are imputed for individuals as the average of indicators of items listed. Take antismoking information index as the example, index = (newspapers or magazines + television + radio + billboard + any other outlets)/5, and then aggregated into the PSU-level index.

Results In Table  2, we list the means of the outcome variables and their potential correlates by country. Warning labels are most often observed in Egypt, exposure to antismoking mass-media messages is highest in Vietnam. Smokers are most frequently exposed to work-site smoking bans in Brazil and Mexico— countries with relatively high quit rates, and smokers’ exposure to tobacco advertising and promotion is greatest in Bangladesh, Russia, and the Philippines—countries with relatively low quit rates. Although the associations between recent quitting and its correlates can vary by country, several patterns emerge (Table 3). Men and those older than 24 years are less likely to be former smokers in all countries. Rural smokers and those with more education are more likely to quit in most countries. The association between wealth and quitting varies across countries, with greater wealth associated with increased quitting in Bangladesh, Brazil, Uruguay, Russia, and Ukraine but less

quitting in India and Turkey. This finding suggests that wealthier smokers in a majority of LMICs may have more incentives to quit, which has been shown theoretically in the health capital model developed by Grossman (1972) and empirically shown by others (Fagan et al., 2007; Siahpush, McNeill, Borland, & Fong, 2006). Given intention to quit, wealthier smokers also tend to have more access to professional services and drugs that help quitting (Kotz & West, 2009). Higher cigarette prices are significantly associated with increased odds of being a recent quitter in the Philippines, and higher bidi prices are significantly associated with increased odds of being a recent quitter in Bangladesh, with a marginally significant association seen for India. Greater exposure to mass-media antismoking information is significantly associated with increased odds of quitting in Poland. Greater exposure to tobacco advertising and promotion is significantly associated with less quitting in Bangladesh and the Philippines. And greater awareness of warning labels is associated with higher quit rates in Egypt. Mixed results are obtained for

S47

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

PSU-level variables

S48

% High school

% Secondary school

% Primary school

% Age 65+

% Age 40–64

% Age 25–39

% Outdoor occupation

% Indoor occupation

Wealth

% Male

% Rural

% Quit in past 12 months % Quit attempts in past 12 months Household size

11.5 (31.9) 41.8 (49.3) 3.26 (1.79) 17.9 (38.4) 56.8 (49.5) 8.57 (2.38) 22.6 (41.8) 36.6 (48.2) 33.1 (47.1) 44.7 (49.7) 8.91 (28.5) 18.9 (39.2) 39.3 (48.8) 19.4 (39.6)

BR 15.9 (36.5) 46.2 (49.9) 4.53 (2.13) 36.5 (48.1) 75.8 (42.8) 9.63 (2.72) 22.3 (41.6) 47.4 (49.9) 36.0 (48.0) 32.1 (46.7) 7.02 (25.6) 25.2 (43.4) 29.6 (45.6) 14.6 (35.3)

MX 11.4 (31.8) 44.0 (49.7) 3.15 (1.86) 33.7 (47.3) 57.2 (49.5) 11.3 (2.94) 30.2 (45.9) 33.3 (47.1) 33.5 (47.2) 42.1 (49.4) 8.46 (27.8) 43.3 (49.6) 21.5 (41.1) 15.3 (36.0)

UY 3.10 (17.3) 30.5 (46.0) 5.14 (2.37) 68.3 (46.5) 88.4 (32.1) 1.22 (0.82) 15.6 (36.2) 56.7 (49.5) 37.2 (48.3) 45.3 (49.8) 9.25 (29.0) 29.5 (45.6) 25.6 (43.6) 6.63 (24.9)

IN 4.29 (20.3) 47.7 (50.0) 4.49 (1.87) 52.4 (50.0) 96.5 (18.3) 0.53 (0.26) 16.8 (37.4) 74.7 (43.5) 42.8 (49.5) 40.0 (49.0) 6.34 (24.4) 26.6 (44.2) 20.4 (40.3) 2.83 (16.6)

BD 5.34 (22.5) 47.4 (49.9) 3.53 (1.69) 44.9 (49.7) 91.0 (28.7) 6.55 (1.75) 17.0 (37.6) 58.3 (49.3) 30.4 (46.0) 48.1 (50.0) 11.9 (32.4) 53.8 (49.8) 17.4 (37.9) 14.8 (35.5)

TH 4.52 (20.8) 30.9 (46.2) 2.78 (1.30) 62.1 (48.5) 93.8 (24.1) 4.26 (1.58) 24.7 (43.2) 41.4 (49.3) 21.2 (40.9) 58.2 (49.3) 16.7 (37.3) 27.3 (44.5) 39.1 (48.8) 16.8 (37.4)

CN 7.57 (26.5) 51.7 (50.0) 3.89 (1.64) 51.9 (50.0) 96.0 (19.7) 2.82 (1.25) 20.7 (40.5) 50.6 (50.0) 35.0 (47.7) 47.5 (49.9) 8.22 (27.5) 26.8 (44.3) 26.0 (43.9) 25.6 (43.6)

VN 6.77 (25.1) 46.5 (49.9) 5.00 (2.35) 60.2 (49.0) 82.8 (37.7) 1.85 (1.08) 11.3 (31.7) 65.9 (47.4) 39.4 (48.9) 37.5 (48.4) 7.68 (26.6) 40.5 (49.1) 16.9 (37.5) 19.7 (39.8)

PH 7.43 (26.2) 30.4 (46.0) 3.35 (1.61) 47.0 (49.9) 58.8 (49.2) 16.2 (2.74) 34.1 (47.4) 23.3 (42.3) 33.0 (47.0) 50.8 (50.0) 6.09 (23.9) 13.6 (34.2) 35.8 (47.9) 36.1 (48.0)

PL

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

Table 2.  Descriptive Statistics TR 9.85 (29.8) 41.1 (49.2) 4.11 (2.06) 44.1 (49.7) 74.3 (43.7) 11.5 (1.55) 29.1 (45.4) 33.0 (47.0) 42.6 (49.5) 41.1 (49.2) 4.81 (21.4) 51.6 (50.0) 10.4 (30.5) 20.2 (40.2)

RU 5.11 (22.0) 29.2 (45.5) 3.04 (1.39) 46.0 (49.8) 78.3 (41.2) 11.1 (2.33) 41.5 (49.3) 32.2 (46.7) 33.8 (47.3) 44.1 (49.7) 6.02 (23.8) 2.17 (14.6) 6.99 (25.5) 69.3 (46.1)

8.89 (28.5) 34.8 (47.6) 3.08 (1.38) 47.2 (49.9) 84.0 (36.7) 4.84 (1.26) 27.7 (44.7) 28.0 (44.9) 37.4 (48.4) 41.8 (49.3) 7.72 (26.7) 6.45 (24.5) 35.9 (47.9) 41.5 (49.2)

UA

5.50 (22.8) 38.0 (48.6) 4.80 (2.42) 41.1 (49.2) 98.5 (12.3) 4.81 (1.01) 22.8 (42.0) 59.5 (49.1) 39.1 (48.8) 43.4 (49.6) 6.80 (25.2) 18.1 (38.5) 11.2 (31.5) 8.14 (27.4)

EG

(Continued)

6.81 (25.2) 38.3 (48.6) 3.98 (2.13) 48.0 (50.0) 81.4 (38.9) 6.05 (4.61) 22.9 (42.0) 47.1 (49.9) 34.9 (47.7) 44.9 (49.7) 8.73 (28.2) 26.8 (44.3) 24.5 (43.0) 20.8 (40.6)

All

Quitters in low- and middle-income countries

10.0 (30.0) 5.38 (0.55) 0.85 (0.41) 2.78 (0.92) 5.95 (1.41) 1.56 (0.35) 3.90 (1.64) – 2,164

7,915

MX

6.85 (25.3) 5.80 (0.74) 0.32 (0.20) 3.74 (1.06) 7.92 (0.87) 1.58 (0.21) 1.44 (0.65) –

BR

1,573

2.69 (0.94) –

6.17 (24.1) 4.72 (0.69) 0.61 (0.35) 1.61 (0.66) 8.80 (0.94) –

UY 7.59 (26.5) 3.45 (1.94) 0.52 (0.69) 0.85 (1.05) 6.11 (2.16) 1.07 (0.63) 2.79 (1.76) 0.73 (0.62) 11,967

IN 4.16 (20.0) 2.50 (1.19) 2.20 (1.05) 1.88 (1.27) 5.81 (1.53) 0.81 (0.57) 1.26 (0.31) 0.29 (0.15) 2,333

BD

5,184

9.48 (29.3) 5.86 (0.64) 0.38 (0.23) 0.24 (0.15) 8.65 (0.56) 1.46 (0.21) 1.83 (0.49) –

TH

4,200

8.81 (28.3) 2.85 (1.48) 0.27 (0.34) 0.45 (0.41) 6.25 (1.58) 0.76 (0.31) 1.72 (2.75) –

CN

2,445

0.37 (6.06) 6.17 (1.52) 0.28 (0.36) 0.47 (0.56) 8.70 (1.17) 0.80 (0.52) 2.50 (1.17) –

VN

2,970

19.0 (39.3) 4.75 (2.02) 1.58 (1.32) 3.73 (1.82) 8.33 (1.64) 1.29 (0.60) 0.85 (0.60) –

PH

2,610

14.3 (35.0) 3.76 (1.53) 0.44 (0.39) 0.60 (0.46) 8.54 (1.16) 1.32 (0.30) 4.36 (1.01) –

PL

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

5,066

21.3 (41.0) 2.75 (1.33) 1.20 (0.99) 2.80 (1.65) 8.48 (1.59) 1.15 (0.26) 1.26 (0.82) –

RU

2,996

8.71 (28.2) 4.08 (1.08) 0.26 (0.29) 0.22 (0.24) 8.15 (1.30) 1.27 (0.50) 3.29 (0.79) –

TR

2,631

16.0 (36.6) 2.90 (1.48) 0.61 (0.74) 1.61 (1.21) 7.97 (1.90) 1.30 (0.39) 1.43 (0.18) –

UA

4,397

37.3 (48.4) 2.87 (0.42) 0.21 (0.14) 0.33 (0.17) 9.75 (0.20) 0.84 (0.24) 4.61 (6.75) –

EG

Note. In the column headers, BR, MX, UY, IN, BD, TH, CN, VN, PH, PL, RU, TR, UA, and EG represent Brazil, Mexico, Uruguay, India, Bangladesh, Thailand, China, Vietnam, the Philippines, Poland, Russia, Turkey, Ukraine, and Egypt, respectively. The samples are restricted to current smokers and past year quitters. For quit attempt rates, the samples for denominators are current smokers. Standard deviation is in the parenthesis.

N

Bidi prices

Work-site smoking ban index Cigarette prices

Antismoking information index Tobacco promotion index Tobacco advertising index Warning label index

% College or greater

Table 2. Continued

12.1 (32.6) 4.10 (1.85) 0.61 (0.78) 1.54 (1.65) 7.64 (1.92) 1.18 (0.50) 2.35 (2.50) 0.66 (0.29) 58,451

All

Nicotine & Tobacco Research, Volume 16, Supplement 1 (January 2014)

S49

Household size Rural Male Wealth Indoor occupation Outdoor occupation Age  25–39  40–64  65+ Education  Primary

Household size Rural Male Wealth Indoor occupation Outdoor occupation Age  25–39  40–64  65+ Education  Primary  Secondary   High school   College or above PSU-level indices   Antismoking information   Tobacco promotion   Tobacco advertising   Warning label   Work ban* indoor occupation   Work ban* other occupation   Cigarette price   Bidi price 0.47–0.70 0.33–0.51 0.33–0.64 0.69–1.18 0.67–0.99 0.71–1.11 0.82–1.57 0.87–1.07 0.57–1.22 0.96–1.11 0.97–1.17 0.25–1.02 0.93–2.09 0.87–1.10 –

0.57*** 0.41*** 0.46*** 0.90 0.82** 0.89 1.14 0.97 0.84 1.03 1.07 0.51* 1.39 0.98 –

0.60–1.34 0.24–0.62 0.31–1.46

0.90 0.38*** 0.68 1.00

0.88–1.13 0.67–1.53 0.72–1.28 0.97–1.11 0.16–4.19 0.46–1.40

1.00 1.02 0.96 1.04 0.83 0.80

PLa (N = 2,604)

0.93–1.01 0.92–1.43 0.71–0.97 1.01–1.09 1.11–13.4 0.60–0.86

(N = 2,604)

95% CI

0.97 1.14 0.83** 1.05** 3.85** 0.72***

PLa

0.46*** 0.41*** 0.49**

0.89** 1.30* 0.73** 1.07* 1.50 1.01

1.00

0.64–1.15 0.49–1.29 0.89–1.38 0.85–1.08 0.39–3.11 0.84–1.81 0.95–1.08 –

0.80–1.79 0.83–1.95 0.65–1.81 0.63–2.20

0.46–0.90 0.34–0.66 0.34–1.08

1.01–1.13 0.82–1.57 0.49–0.97 0.92–1.04 0.21–6.06 0.63–1.24

(N = 5,059)

95% CI

0.34–0.63 0.30–0.56 0.25–0.96

0.81–0.99 0.98–1.73 0.54–0.97 1.00–1.14 0.43–5.26 0.68–1.51

RUa (N = 5,059)

0.86 0.79 1.11 0.96 1.10 1.23 1.01 –

1.20 1.27 1.09 1.18

0.64*** 0.47*** 0.61*

1.07** 1.13 0.69** 0.98 1.12 0.88

RUa

OR

0.72–1.25 0.55–1.48 0.77–1.36 0.70–1.01 – – 0.83–1.32 –

1.05–3.84 0.60–3.03 1.02–5.49 0.34–2.98

0.37–0.90 0.32–0.77 0.24–0.95

0.84–1.02 1.09–2.41 0.48–1.00 1.02–1.16 0.54–1.22 0.56–1.36

0.68* 0.87 1.09

0.95 1.39* 0.66*** 0.93* 0.49 0.54***

0.67–1.66

0.45–1.03 0.59–1.30 0.61–1.96

0.89–1.02 1.00–1.94 0.49–0.90 0.87–1.01 0.18–1.39 0.37–0.78

TR (N = 2,996)

0.95 0.90 1.03 0.84* – – 1.04 –

2.01** 1.35 2.37** 1.01

0.58** 0.49*** 0.48**

0.93 1.58** 0.70** 1.09** 0.81 0.87

1.06

95% CI

TR (N = 2,996)

OR

95% CI

0.45–1.04 0.35–0.89 0.48–1.78 1.06–2.03 1.27–2.56 0.75–2.11 1.56–3.52 0.91–1.08 0.91–1.45 0.80–1.11 0.87–1.01 0.71–2.34 0.97–1.42 0.87–1.02 0.79–1.46

0.69* 0.56** 0.92 1.46** 1.80*** 1.26 2.34*** 0.99 1.15 0.94 0.94 1.29 1.17 0.95 1.07

0.88–1.05 0.70–2.14 0.40–0.76 1.09–1.41 0.28–2.52 0.47–0.98 0.39–0.94 0.31–0.72 0.24–0.94 0.04–8.49

0.96 1.23 0.55*** 1.24*** 0.85 0.68** 0.61** 0.47*** 0.48** 0.60

UA (N = 2,623)

0.95–1.05 0.70–1.18 0.58–1.55 0.61–1.02 0.30–1.35 0.47–0.92

1.00 0.90 0.95 0.79* 0.63 0.66**

UA (N = 2,623)

OR

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

S50 OR

Table 3.  Country-Specific Models of Recent Quitting (Quitting in the Past 12 Months)

0.28*** 0.52** 0.87

1.08* 0.81 1.02 3.26** 0.41 0.47**

0.96 1.35** 0.89 1.04 1.04 0.57***

0.92

0.90 1.25 1.20

0.96–1.49 0.65–1.02 0.80–1.22 0.82–1.20 0.47–3.78 0.51–1.18 0.23–3.08 2.05–12.5

0.44–1.29 0.40–1.38 0.26–2.75 0.47–3.45

0.15–0.52 0.29–0.95 0.40–1.89

0.99–1.18 0.40–1.65 0.36–2.90 1.17–9.12 0.12–1.38 0.24–0.91

(Continued)

0.62–1.36

0.57–1.41 0.78–1.99 0.65–2.24

0.91–1.02 1.05–1.74 0.35–2.29 0.92–1.18 0.33–3.30 0.40–0.80

EG (N = 4,395)

1.20 0.81* 0.99 0.99 1.34 0.78 0.84 5.06***

0.76 0.75 0.85 1.28

95%CI

EG (N = 4,395)

OR

Quitters in low- and middle-income countries

S51

1.13–1.42 0.58–1.19 0.81–1.55 0.89–1.22 0.22–2.03 0.41–1.06 0.95–1.15

1.26*** 0.84 1.12 1.04 0.67 0.65* 1.04

0.87–1.01 0.89–1.56 0.43–1.02 0.98–1.13 0.29–14.6 0.42–0.74 0.38–0.82 0.45–0.95 0.43–1.17 0.60–2.14 0.56–2.31 0.63–2.72 0.66–3.32 0.96–1.39 0.91–2.45 0.46–2.46 0.77–1.08 0.22–2.22 0.72–2.73 0.76–1.36

0.94 1.18 0.66* 1.05 2.07 0.56*** 0.55*** 0.66** 0.71 1.13 1.14 1.31 1.48 1.15 1.49 1.07 0.91 0.70 1.41 1.01

TH (N = 5,184)

0.51–1.77 0.63–2.65 0.61–3.00

(N = 2,604)

95% CI

0.95 1.29 1.35

PLa

OR

1.11* 1.15 0.71 0.98 0.16** 0.61***

1.02 1.08 0.72 0.99 6.37** 1.90* 1.02

0.85 0.64 0.97 0.98

0.71 0.60 1.06

0.96–1.19 1.05–1.39 0.86–1.05 0.96–1.17 0.71–2.69 0.82–3.46 0.72–1.10

0.78–15.7 0.58–10.6 0.66–12.8

(N = 5,059)

95% CI

0.83–1.24 0.59–1.97 0.45–1.17 0.85–1.15 1.28–31.7 0.94–3.81 0.99–1.04

0.54–1.36 0.35–1.17 0.48–1.95 0.40–2.40

0.35–1.45 0.31–1.18 0.49–2.31

0.98–1.26 0.73–1.79 0.43–1.18 0.86–1.12 0.03–0.87 0.43–0.86

CN (N = 4,197)

1.07 1.21*** 0.95 1.06 1.38 1.68 0.89

3.49 2.49 2.92

RUa

OR

1.16 1.34 1.64* 1.65

0.60** 0.57** 0.67

1.00 0.82 0.89 1.14 1.55 1.07 1.02

0.82–1.07 0.56–1.38 0.71–2.47 0.88–1.09 0.79–2.91 0.73–1.41 0.96–1.33

0.55–1.84 0.58–1.75 0.69–2.42

0.90–1.13 0.50–1.37 0.65–1.24 0.97–1.34 0.79–3.05 0.78–1.49 0.87–1.19

0.67–1.99 0.80–2.27 0.93–2.91 0.20–13.6

0.37–0.97 0.36–0.89 0.36–1.26

0.86–1.06 0.79–1.68 0.24–1.01 0.93–1.28 0.28–1.39 0.50–0.98

VN (N = 2,444) 0.95 1.15 0.49* 1.09 0.62 0.70**

0.94 0.88 1.32 0.98 1.52 1.01 1.13

1.01 1.01 1.29

95% CI

TR (N = 2,996)

OR

1.00 0.81*** 1.09 0.98 1.35 0.93 1.20***

1.78 1.54 1.49 2.16

0.97–1.15 0.74–1.14 0.80–1.09 0.95–1.12 0.52–1.76 0.74–1.60 0.19–3.80

0.05–6.72 0.04–6.34 0.04–7.11

0.92–1.09 0.71–0.91 0.98–1.20 0.87–1.09 0.67–2.72 0.70–1.23 1.04–1.38

0.69–4.60 0.55–4.29 0.53–4.20 0.76–6.10

0.37–0.87 0.51–1.19 0.64–2.15

0.95–1.09 0.72–1.42 0.48–1.11 0.93–1.29 0.16–1.88 0.50–1.07

PH (N = 2,969)

0.56*** 0.78 1.18

1.02 1.01 0.73 1.10 0.55 0.73

1.05 0.92 0.93 1.03 0.95 1.09 0.86

0.57 0.50 0.53

95% CI

UA (N = 2,623)

OR

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

1.01 0.75 1.50 3.20*** 0.68 1.20 0.98

0.83 0.71 1.23

95%CI

0.74–1.39 0.30–1.88 0.62–3.66 1.53–6.68 0.22–2.08 0.57–2.51 0.95–1.01

0.48–1.42 0.41–1.24 0.86–1.75

EG (N = 4,395)

OR

Note. aThere are very few respondents in Poland and Russia who have not received formal education. Therefore the base category of education indicators is primary education for Poland and Russia. OR = odds ratio; CI = confidence interval. *.05 < p ≤ .1, **.01 < p ≤ .05, ***p ≤ .01.

Household size Rural Male Wealth Indoor occupation Outdoor occupation Age  25–39  40–64  65+ Education  Primary  Secondary   High school   College or above PSU-level indices   Antismoking information   Tobacco promotion   Tobacco advertising   Warning label   Work ban* indoor occupation   Work ban* other occupation   Cigarette price

 Secondary   High school   College or above PSU-level indices   Antismoking information   Tobacco promotion   Tobacco advertising   Warning label   Work ban* indoor occupation   Work ban* other occupation Cigarette price

Table 3. Continued

Nicotine & Tobacco Research, Volume 16, Supplement 1 (January 2014)

0.93–1.07 1.02–1.45 0.72–0.87 0.99–1.09 0.85–6.82 0.67–0.89 0.55–0.63 0.38–0.49 0.41–0.56 0.80–1.33 0.70–1.16 0.74–1.33 1.04–1.16 0.92–1.01 0.74–0.82 1.02–1.08 0.88–1.09 0.35–1.24 1.17–1.43 0.99–1.03 –

1.00 1.21** 0.79*** 1.04 2.41* 0.77*** 0.59*** 0.43*** 0.48*** 1.03 0.90 0.99 1.10*** 0.96* 0.78*** 1.05*** 0.98 0.66 1.29*** 1.01 –

95% CI

0.99 1.06 0.96* 0.94*** 1.20*** 1.12* 0.94*** 1.07***

1.27** 1.43*** 1.35*** 1.87***

0.58*** 0.60*** 0.83**

0.99 1.05 0.84 0.99 0.66*** 0.60***

OR

0.93–1.05 0.89–1.27 0.92–1.00 0.93–0.95 1.06–1.35 1.00–1.25 0.92–0.97 1.04–1.11

1.02–1.58 1.09–1.88 1.15–1.60 1.53–2.28

0.43–0.79 0.53–0.68 0.70–0.98

0.95–1.03 0.91–1.20 0.63–1.13 0.88–1.12 0.56–0.78 0.52–0.70

95% CI

1.01 0.83*** 1.04 1.00 1.85* 1.09 1.03*** –

1.22 1.07 1.38 1.54*

0.61*** 0.65*** 1.08

1.01 1.12* 0.68*** 1.04 0.43** 0.68***

OR

1.00–1.02 0.79–0.87 0.94–1.16 0.94–1.07 0.93–3.68 0.84–1.41 1.01–1.05 –

0.86–1.73 0.62–1.82 0.88–2.16 0.98–2.42

0.56–0.67 0.53–0.81 0.84–1.38

0.96–1.07 1.00–1.26 0.58–0.81 0.97–1.10 0.19–0.99 0.61–0.75

95% CI

Asia Pacific (N = 9,615) 95% CI 0.93–0.98 1.16–1.41 0.59–0.88 1.01–1.10 0.49–1.32 0.61–0.97 0.46–0.76 0.32–0.69 0.39–0.87 – 0.82–1.16 0.83–1.11 1.00–1.24 0.96–1.21 0.90–1.30 0.92–1.04 1.00–1.06 0.90–1.49 0.87–1.25 0.90–1.09 –

OR 0.95*** 1.28*** 0.72*** 1.05** 0.80 0.77** 0.59*** 0.47*** 0.58*** – 0.98 0.96 1.11* 1.08 1.08 0.98 1.03* 1.16 1.05 0.99 –

Europe (N = 13,303)

1.03 0.98 1.01 0.99 1.18 1.13*** 0.99 –

1.04 1.01 1.07 1.26**

0.58*** 0.52*** 0.68***

0.99 1.18*** 0.78*** 1.04** 0.78 0.70***

OR

0.98–1.08 0.85–1.12 0.97–1.04 0.95–1.03 0.94–1.46 1.04–1.22 0.97–1.01 –

0.86–1.26 0.80–1.27 0.87–1.32 1.03–1.52

0.53–0.64 0.43–0.62 0.53–0.87

0.96–1.02 1.10–1.26 0.71–0.86 1.00–1.07 0.57–1.09 0.64–0.77

95% CI

All (N = 58,451)

Note. All regressions also control for country fixed effects. In the last two columns, the sample includes all 14 countries. The pooled sample of America includes Mexico, Brazil, and Uruguay. The pooled sample of South East Asia includes India, Bangladesh, and Thailand, where the bidi price for Thailand is replaced by the mean of the bidi price for India and Bangladesh. The pooled sample of Asia Pacific includes China, Vietnam, and the Philippines. The pooled sample of Europe includes Poland, Russia, Ukraine, and Turkey. The information on indoor work-site smoking policy in Uruguay is not available, and its work-site smoking ban index is replaced by the mean index of the other countries in the pooled models of America and all countries. CI = confidence interval; OR = odds ratio. *.05 < p ≤ .1, **.01 < p ≤ .05, ***p ≤ .01.

Household size Rural Male Wealth Indoor occupation Outdoor occupation Age  25–39  40–64  65+ Education  Primary  Secondary   High school   College or above PSU-level indices   Antismoking information   Tobacco promotion   Tobacco advertising   Warning label   Work ban* indoor occupation   Work ban* other occupation Cigarette price Bidi price

OR

Southeast Asia (N = 19,484)

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

S52 Latin America (N = 11,652)

Table 4.  Pooled Models of Recent Quitting (Quitting in the Past 12 Months)

Quitters in low- and middle-income countries

0.96–1.03 0.81–1.29 0.69–0.82 0.99–1.00 1.46–2.19 0.95–1.05 0.96–1.06 0.73–0.95 0.57–0.78 0.84–1.44 0.96–1.26 0.75–1.34 0.66–1.00 1.00–1.11 0.95–1.01 0.98–1.04 0.96–1.02 0.63–0.89 1.09–1.13 0.92–1.00 –

0.99 1.02 0.75*** 1.00 1.79*** 1.00 1.01 0.83*** 0.67*** 1.10 1.10 1.00 0.81* 1.05* 0.98 1.01 0.99 0.75*** 1.11*** 0.96* –

95% CI

1.01 1.09* 1.02 1.04*** 1.09 1.21*** 0.99*** 1.17***

1.05 1.30*** 1.22*** 1.25***

0.89** 0.90 0.88

1.01 0.95* 0.87*** 0.99 1.33 1.18***

OR

0.94–1.07 0.98–1.20 0.99–1.05 1.02–1.07 0.90–1.32 1.08–1.37 0.98–1.00 1.11–1.22

0.98–1.13 1.17–1.43 1.12–1.33 1.11–1.40

0.80–1.00 0.78–1.03 0.64–1.23

0.97–1.06 0.90–1.00 0.79–0.96 0.89–1.10 0.93–1.88 1.08–1.28

95% CI

Southeast Asia (N = 18,606)

1.09*** 1.04 1.02 1.00 1.02 0.98 1.01 –

1.16* 1.50*** 1.53*** 1.39***

1.01 0.91 0.82***

0.99 1.26*** 0.78** 1.02 0.97 0.85***

OR

1.07–1.12 0.99–1.10 0.94–1.10 0.96–1.04 0.72–1.45 0.87–1.10 0.99–1.03 –

0.98–1.38 1.24–1.81 1.18–1.98 1.21–1.58

0.84–1.22 0.67–1.23 0.71–0.94

0.96–1.02 1.13–1.39 0.62–0.99 0.95–1.10 0.68–1.38 0.79–0.91

95% CI

Asia Pacific (N = 6,607)

1.20*** 1.03** 0.97 1.05** 0.81 0.96 1.04*** –

– 1.01 0.97 0.95

0.75*** 0.65*** 0.55***

1.01 1.19*** 0.86** 1.03** 1.13 0.95

1.11–1.30 1.00–1.05 0.91–1.04 1.01–1.09 0.62–1.05 0.89–1.05 1.04–1.05 –

– 0.99–1.04 0.80–1.17 0.74–1.23

0.68–0.82 0.53–0.80 0.41–0.73

0.98–1.04 1.11–1.28 0.75–0.99 1.00–1.05 0.82–1.55 0.86–1.04

95% CI

Europe (N = 12,307) OR

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

Note. See Note for Table 4. OR = odds ratio; CI = confidence interval. *.05 < p ≤ .1, **0.01 < p ≤ .05, ***p ≤ .01.

Household size Rural Male Wealth Indoor occupation Outdoor occupation Age  25–39  40–64  65+ Education  Primary  Secondary   High school   College or above PSU-level indices   Antismoking information   Tobacco promotion   Tobacco advertising   Warning label   Work ban* indoor occupation   Work ban* other occupation Cigarette price Bidi price

OR

Latin America (N = 10,215)

Table 5.  Pooled Models of Quit Attempts (Unsuccessful Quit Attempt in the Past 12 Months)

1.08** 1.05** 1.01 1.03*** 0.97 1.11* 1.01*** –

1.09** 1.23*** 1.17*** 1.12**

0.89*** 0.81*** 0.75***

1.00 1.10** 0.83*** 1.00 1.18** 1.00

1.00–1.17 1.00–1.10 0.98–1.04 1.01–1.05 0.83–1.12 0.99–1.26 1.00–1.02 –

1.02–1.17 1.14–1.32 1.06–1.30 1.00–1.25

0.82–0.97 0.72–0.92 0.61–0.92

0.98–1.02 1.01–1.19 0.77–0.89 0.98–1.02 1.00–1.39 0.89–1.13

95% CI

All (N = 51,890) OR

Nicotine & Tobacco Research, Volume 16, Supplement 1 (January 2014)

S53

Quitters in low- and middle-income countries

Discussion In this study, we use GATS data from 14 countries to describe the factors associated with quitting and quit attempts. We find that living in rural areas, having more education, and being wealthier are factors associated with higher odds of being a quitter and trying to quit. Men are less likely to have quit or tried to quit than women. Greater exposure to work-site smoking bans is associated with higher odds of recent quitting. Although higher cigarette prices are associated with higher probability of quit attempts, higher bidi prices are associated with higher probabilities of both quitting and quit attempts in Southeast Asian countries where bidi use is common. Greater exposure to work-site bans is strongly associated with quitting in China where smoking at indoor work-sites is prevalent. Our estimates also call attention to the potential influence of tobacco marketing in the Asia Pacific and Latin American regions where greater exposure to tobacco promotion is linked to reduced likelihood of quitting. Our findings are in line with the existing limited literature that investigates quitting and quit attempts in LMICs. For example, we consistently find that quitting and quit attempts

S54

in high tobacco-using LMICs are low, which has been documented in a series of reports using GATS (http://nccd.cdc. gov/gtssdata/Ancillary/DataReports) and reports from the Bloomberg Global Initiative to Reduce Tobacco Use (www. tobaccofreeunion.org/content/en/217). In this study, we further explore how quitting and quit attempts are associated with individual and environmental risk factors. Our findings show that although the associations between these factors and quitting may vary by countries, the results of pooled analyses that take account of unobserved country-specific attributes tend to indicate that tobacco control polices such as work-site smoking bans, warning labels, and antismoking media messaging can be linked to either quitting or quit attempts. Meanwhile, we have observed that not all LMICs are at the same stage of employing these tobacco control policies or are not applying them at the same level so that there are substantial differences across countries in exposure to them. The potential of these policies in encouraging quitting may be especially relevant to policy makers in countries where they do not appear to reach enough smokers. There are some limitations in this analysis. We use crosssectional data from 14 countries to study cessation. The cessation measures, prices, and indices for policy and marketing exposures are constructed using self-reported information. In addition, these prices and indices are contemporaneous measures; while most previous literature has shown that it is the change of prices or policies over time that drives quitting, we cannot estimate the effects of changes over time in this study. However, given that most tobacco policies are recently adopted in LMICs, we find their associations with quitting to be significant and strong even when these policies are contemporaneously measured. This study takes the initial steps in investigating the associations between determinants and quitting across LMICs, but future research that employs longitudinal surveys in many countries is needed to better understand the effectiveness and cost-effectiveness of tobacco control interventions in LMICs.

Funding Funding for the Global Adult Tobacco Survey (GATS) is provided by the Bloomberg Initiative to Reduce Tobacco Use, a program of Bloomberg Philanthropies. Governments of Brazil and India contributed to GATS implementation in their respective countries. The Bill and Melinda Gates Foundation provided additional funding for GATS implementation in China and for analysis.

Declaration of Interests The conclusions in this paper are those of the authors and do not necessarily represent the official position of their affiliated organizations.

Acknowledgments We thank Nahleen Zahra, Pavel Dramski, and William Ridgeway for excellent research assistance. The findings of this study are those of the authors and do not represent the official position of the Centers for Disease Control and Prevention.

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

exposure to indoor work-site smoking bans. The estimates indicate that working in an outdoor occupation is associated with lower odds of quitting for most countries, and working indoors is associated with higher odds of quitting in Brazil and lower odds in China. Exposure to work-site smoking bans is associated with increased odds of quitting for Chinese smokers who work indoors, and the association is particularly strong, which may reflect the social norms around smoking as a way of networking at work in the country. In the pooled models of recent quitting (Table  4), living in rural areas, having higher education, and more wealth are factors associated with higher odds of being a recent quitter, whereas being male, over 24  years old, and working in an outdoor occupation are factors associated with lower odds. Although exposure to work-site smoking bans increases the odds of being a recent quitter, greater exposure to cigarette advertising is associated with lower odds of quitting in the Southeast Asian region. Similarly, greater exposure to tobacco promotion is associated with lower odds of quitting in Latin America and Asia Pacific regions. The warning label index is associated with higher odds of quitting in European countries but lower odds of quitting in Southeast Asia. The associations between individual demographic characteristics and the probability of making a recent quit attempt are quite similar to the associations observed in the models of recent quitting (Table  5). Living in rural areas is associated with higher odds of quit attempts in most regions other than Southeast Asia. Having formal education is associated with higher odds of quit attempts in most regions other than Latin American. Being male and being older than 24 years are associated with lower odds in all models. Greater exposure to antismoking mass-media messages, higher cigarette prices, and greater exposure to warning labels are significantly associated with increased odds of making a recent quit attempt. Although working indoors is associated with higher odds of a quit attempt, the exposure to work-site smoking bans also affects those who do not usually work indoors in increasing quit attempts.

Nicotine & Tobacco Research, Volume 16, Supplement 1 (January 2014)

References

S55

Downloaded from http://ntr.oxfordjournals.org/ at Freie Universitaet Berlin on May 13, 2015

Blecher, E. H., & van Walbeek, C. P. (2004). An international analysis of cigarette affordability. Tobacco Control, 13, 339– 346. doi:10.1136/tc.2003.006726 Blecher, E. H., & van Walbeek, C. P. (2009). Cigarette affordability trends: An update and some methodological comments. Tobacco Control, 18, 167–175. doi:10.1136/tc.2008.026682 Centers for Disease Control and Prevention. (2012) Increases in Quitline Calls and Smoking Cessation Website Visitors During a National Tobacco Education Campaign—March 19–June 10. Morbidity and Mortality Weekly Report, 61, 667–670. Retrieved from http://www.cdc.gov/mmwr/pdf/ wk/mm6134.pdf Chaloupka, F. J., & Warner, K. E. (2000). The economics of smoking. In A. J. Culyer & J. P. Newhouse (Eds.), Handbook of health economics (pp. 1539–1627). Amsterdam: Elsevier Science, North-Holland. doi:10.1016/S1574-0064(00)80042–6 Chaloupka F. J., Kostova, D., & Shang, C. (2013). Cigarette excise tax structure and cigarette prices—Evidence from the Global Adult Tobacco Survey and the U.S. National Adult Tobacco Survey. Nicotine & Tobacco Research. doi:10.1093/ntr/ntt121 Chaloupka, F. J., Straif, K., & Leon, M. E. (2011). Effectiveness of tax and price policies in tobacco control. Brief report. Tobacco Control, 20, 235–238. doi:10.1136/tc.2010.039982 Davis, K. C., Nonnemaker, J. M., Farrelly, M. C., & Niederdeppe, J. (2011). Exploring differences in smokers’ perceptions of the effectiveness of cessation media messages. Tobacco Control, 20, 26–33. doi:10.1136/tc.2009.035568 DeCicca, P., Kenkel, D., & Mathios, A. (2008). Cigarette taxes and the transition from youth to adult smoking: Smoking initiation, cessation, and participation. Journal of Health Economics, 27, 904–917. doi:10.1016/j.jhealeco.2008.02.008 Fagan, P., Augustson, E., Backinger, C. L., O’Connell, M. E., Vollinger, R. E. Jr, Kaufman, A., & Gibson, J. T. (2007). Quit attempts and intention to quit cigarette smoking among young adults in the United States. American Journal of Public Health, 97, 1412–1420. Grossman, M. (1972). The demand for health: A  theoretical and empirical investigation. Columbia: University Press for the National Bureau of Economic Research. Guindon, G. E., Perucic, A. M., & Boisclair, D. (2003). Higher tobacco prices and taxes in South-east Asia: An effective tool to reduce tobacco use, save lives and increase government revenue. HNP Discussion Paper Economics of Tobacco Control Paper No 11. The World Bank. Retrived from http:// siteresources.worldbank.org/ Hammond, D., Fong, G. T., McNeill, A., Borland, R., & Cummings, K. M. (2006). Effectiveness of cigarette warning labels in informing smokers about the risks of smoking: Findings from the International Tobacco Control (ITC) Four Country Survey. Tobacco Control, 15(Suppl. 3), iii19–iii25. doi:10.1136/tc.2005.012294 IARC Handbooks of Cancer Prevention, Tobacco Control. (2008). Methods for evaluating tobacco control policies (Vol. 12). Lyon, France: IARC. Jha, P., Jacob, B., Gajalakshmi, V., Gupta, P. C., Dhingra, N., Kumar, R., . . . Peto, R.; RGI-CGHR Investigators. (2008). A nationally representative case-control study of smoking

and death in India. New England Journal of Medicine, 358, 1137–1147. doi:10.1056/NEJMsa0707719 King, B., Yong, H., Borland, R., Omar, M., Ahmad, A. A., Sirirassamee, B., … Thrasher, J. (2010). Malaysian and Thai smokers’ beliefs about the harmfulness of ‘light’ and menthol cigarettes. Tobacco Control, 19, 444–450. doi:10.1136/ tc.2009.034256 Kostova, D. (2012). A (nearly) global look at the dynamics of youth smoking initiation and cessation: The role of cigarette prices. Applied Economics, 45, 3943–3951. Kostova, D., Chaloupka, F. J., Yurekli, A., Ross, H., Cherukupalli, R., Andes, L., & Asma, S. (2012). A crosscountry analysis of cigarette prices and affordability: Evidence from the Global Adult Tobacco Survey. Tobacco Control. doi:10.1136/tobaccocontrol-2011–050413 Kostova, D., Ross, H., Blecher, E., & Markowitz, S. (2011). Is youth smoking responsive to cigarette prices? Evidence from low- and middle-income countries. Tobacco Control, 20, 419–424. doi:10.1080/00036846.2012.736947 Kotz, D., & West, R. (2009). Explaining the social gradient in smoking cessation: It’s not in the trying, but in the succeeding. Tobacco Control, 18, 43–46. doi:10.1136/ tc.2008.025981 Rani, M., Bonu, S., Jha, P., Nguyen, S. N., & Jamjoum, L. (2003). Tobacco use in India: Prevalence and predictors of smoking and chewing in a national cross sectional household survey. Tobacco Control, 12, e4. doi:10.1136/tc.12.4.e4 Ranson, M. K., Jha, P., Chaloupka, F. J., & Nguyen, S. N. (2002). Global and regional estimates of the effectiveness and cost-effectiveness of price increases and other tobacco control policies. Nicotine & Tobacco Research, 4, 311–319. doi:10.1080/14622200210141000 Ross, H., Kostova, D., Stoklosa, M., & Leon, M. (in press). The impact of cigarette excise taxes on smoking cessation rates from 1994 to 2010 in Poland, Russia, and Ukraine. Nicotine & Tobacco Research. Shang, C., Chaloupka, F. J., Fong, G. T., & Zahra, N. (2013). The distribution of cigarette prices under different tax structures: Findings from the International Tobacco Control Policy Evaluation (ITC) Project. Tobacco Control. doi:10.1136/ tobaccocontrol-2013–050966 Siahpush, M., McNeill, A., Borland, R., & Fong, G. T. (2006). Socioeconomic variations in nicotine dependence, self-efficacy, and intention to quit across four countries: Findings from the International Tobacco Control (ITC) Four Country Survey. Tobacco Control, 15(Suppl. 3), iii71–iii75. doi:10.1136/tc.2004.008763 Tauras, J. A. (2004). Public policy and smoking cessation among young adults in the United States. Health Policy, 68, 321–332. doi:10.1016/j.healthpol.2003.10.007 World Health Organization (2010a). Economics of tobacco toolkit: Economic analysis of demand using data from the Global Adult TobaccoSurvey (GATS) Retrieved from http://whqlibdoc.who. int/publications/2010/9789241500166_eng.pdf World Health Organization (2010b). Guidelines for implementation of Article 14: Guidelines on demand reduction measures concerning tobacco dependence and cessation. Retrived from http://www.who.int/fctc/protocol/guidelines/adopted/ article_14/en/